Skip to content

v1.4.0 "Mnemosyne" — Stable Release

Latest

Choose a tag to compare

@YenFreya YenFreya released this 10 Jun 04:31
· 1 commit to main since this release
ae2e0d2

Stable Release

Mnemosyne is the latest stable release of Odysseus AI Control Center.

This release builds upon the reliability of Athena while simplifying the architecture even further. The experimental watchdog subsystem has been completely removed and replaced with a more predictable startup validation workflow and manual recovery controls.

Mnemosyne focuses on stability, transparency, and maintainability while preserving the one-click experience for managing local Odysseus environments.


🟢 Release Status

Current Stable Release

Mnemosyne is intended to serve as the long-term stable branch of Odysseus AI Control Center.

Development philosophy:

  • Simpler is better
  • Explicit failures are better than hidden recovery
  • Manual control is preferred over aggressive automation
  • Clear logging is preferred over silent behavior

✨ What's New in Mnemosyne

✅ Startup Validation System

Before declaring startup successful, Mnemosyne now performs a final validation phase.

The system verifies:

  • Docker Engine is responding
  • Ollama API is reachable
  • Odysseus Web UI is reachable
  • Required services are operational

If validation fails:

  • Failure is logged
  • User receives a clear error message
  • Startup aborts safely
  • Partial startup states are cleaned up

🔄 Restart Odysseus Feature

New main menu option:

[6] Restart Odysseus

Restart performs:

  1. Safe shutdown sequence
  2. Docker shutdown
  3. WSL shutdown
  4. Cooldown delay
  5. Clean startup sequence
  6. Startup validation

This provides a predictable recovery workflow without requiring users to close and reopen the control center.


📋 Improved Logging

Mnemosyne improves visibility when startup failures occur.

Instead of background recovery attempts:

  • Errors are written directly to logs
  • Failures are visible immediately
  • Users can diagnose problems more easily

❌ Removed in Mnemosyne

Watchdog System Removed

The entire watchdog subsystem has been removed.

Removed components:

  • Watchdog background process
  • watchdog.pid
  • .watchdog_enabled
  • Watchdog startup interception
  • Watchdog monitoring loop
  • Watchdog restart automation
  • Watchdog configuration menu
  • Watchdog status display
  • Watchdog deduplication logic
  • Watchdog crash recovery state machine

Why it was removed

During testing the watchdog introduced:

  • PID tracking complexity
  • Background process synchronization issues
  • Duplicate watchdog launch scenarios
  • Difficult debugging behavior
  • Increased maintenance burden

The watchdog solved a rare problem while introducing significant complexity.

Mnemosyne replaces this with:

  • Startup validation
  • Manual restart controls
  • Clear logging
  • Simpler architecture

Result:

  • Less code
  • Fewer edge cases
  • Easier maintenance
  • More predictable behavior

📸 Main Features

🧠 Environment Detection

Automatically validates:

  • Docker Desktop
  • Docker CLI
  • Ollama
  • WSL
  • curl

before allowing service operations.


🚀 Service Startup

One-click startup for:

  • Docker Desktop
  • Ollama
  • Odysseus Container

with readiness checks and validation.


🛑 Service Shutdown

Controlled shutdown of:

  • Odysseus
  • Ollama
  • Docker Desktop
  • WSL

to reduce stale state issues.


🔄 Service Restart

Dedicated restart workflow for:

  • Docker
  • WSL
  • Ollama
  • Odysseus

without exiting the control center.


📦 Model Management

Displays:

  • Installed Ollama models
  • Storage usage
  • Model inventory

for quick environment inspection.


📊 Health Monitoring

Health Check provides:

  • Odysseus status
  • Ollama status
  • Docker status
  • Storage statistics
  • Data-path validation

📜 Log Management

Built-in tools for:

  • Viewing logs
  • Viewing crash logs
  • Clearing logs

to simplify troubleshooting.


🔄 Project Evolution

🔥 v1.0.0 "Genesis"

Original proof-of-concept release.

Introduced:

  • Core startup workflow
  • Service management
  • Initial monitoring systems

Limitations included:

  • Heavy output suppression
  • Fixed-delay startup logic
  • Limited diagnostics
  • Early watchdog implementation

☀️ v1.2.0 "Icarus"

Large experimental rewrite.

Added:

  • PowerShell helpers
  • Backup workflows
  • Update systems
  • Advanced automation
  • Recovery features

While powerful, Icarus became increasingly complex.


🦉 v1.3.1 "Athena"

Return-to-stability release.

Focused on:

  • Simplicity
  • Reliability
  • Reduced automation
  • Easier maintenance

Athena became the stable baseline.


🧠 v1.4.0 "Mnemosyne"

Refines Athena's philosophy even further.

Added

  • Startup Validation System
  • Restart Odysseus menu option
  • WSL shutdown during restart
  • Improved startup diagnostics
  • Clearer failure handling

Removed

  • Entire watchdog subsystem
  • Background monitoring process
  • PID tracking files
  • Automatic crash recovery
  • Watchdog settings
  • Watchdog health display

Improved

  • Startup reliability
  • Recovery workflow
  • Log transparency
  • Maintainability
  • Code simplicity

📋 Requirements

Required

  • Windows 10 or Windows 11
  • Docker Desktop
  • Ollama
  • Odysseus

Recommended

  • WSL 2 enabled
  • Administrative privileges for certain operations

🤖 AI-Assisted Development Disclosure

This project was created as a learning project using AI-assisted development tools.

Estimated AI contribution:

80–95%

The developer provided:

  • Direction
  • Testing
  • Validation
  • Integration
  • Final design decisions

⚠️ Testing Notice

This project has been tested primarily on the developer's personal system.

Although Mnemosyne has been stable during development, it has not been extensively tested on other hardware or Windows configurations. Compatibility with all environments cannot be guaranteed.

Please report any bugs or compatibility issues through GitHub Issues.


👨‍💻 Author

Created and maintained by Yen The Programmer

Learning software development through experimentation, self-hosting, open-source projects, and AI-assisted workflows.


📌 License

MIT License

You are free to:

  • Use
  • Modify
  • Fork
  • Redistribute
  • Reuse in personal or commercial projects

at your own discretion.